Congestion was shown to hamper organ perfusion, but the exact timing of diuretic initiation during hemodynamic de-escalation in shock is unclear. The aim of this study was to describe the hemodynamic effects of diuretic initiation in the stabilized shock. We performed a monocentric, retrospective analysis, in a cardiovascular medico-surgical ICU. We included consecutive resuscitated adult patients, for whom the clinician decided to introduce loop diuretic treatment for clinical signs of fluid overload. The patients were hemodynamically evaluated at the moment of diuretic introduction and 24h later. Seventy ICU patients were included in this study, with a median duration of ICU stay before diuretic initiation of 2 [1-3]days. 51(73%) patients were classified as congestive (central venous pressure > 12 mmHg). After treatment, the cardiac index increased towards normal values in the congestive group (2.7 ± 0.8L min- 1m- 2 from 2.5 ± 0.8L min- 1m- 2, p = 0.042), but not in the non-congestive group (2.7 ± 0.7L min- 1m- 2 from baseline 2.7 ± 0.8L min- 1m- 2, p = 0.968). A decrease in arterial lactate concentrations was observed in the congestive group (2.1 ± 2 mmol L- 1 vs. 1.3 ± 0.6 mmol L- 1, p < 0.001). The diuretic therapy was associated with an improvement of ventriculo-arterial coupling comparing with baseline values in the congestive group (1.69 ± 1 vs. 1.92 ± 1.5, p = 0.03). The norepinephrine use decreased in congestive patients (p = 0.021), but not in the non-congestive group (p = 0.467). The initiation of diuretics in ICU congestive patients with stabilized shock was associated with improvement of cardiac index, ventriculo-arterial coupling, and tissue perfusion parameter. These effects were not observed in non-congestive patients.